Enterprise Assist: Floating Treatment Wetlands (FTWs) in Water Treatment: Treatment efficiency and potential benefits of activated carbon ...
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
SEACAMS Swansea University
Enterprise Assist:
Floating Treatment Wetlands
(FTWs) in Water Treatment:
Treatment efficiency and
potential benefits of activated
carbon.
I Dodkins & AF Mendzil
March 2014
Prepared for: FROG Environmental Ltd, Ban y Berllan,
Llansadwrn, Llanwrda, SA19 8NA.
Sustainable Expansion of the Applied Coastal
And Marine Sectors (SEACAMS)
Prifysgol Abertawe/Swansea University
Abertawe/Swansea
Cymru/Wales
SA2 8PP
www.seacams.ac.uk
Ebost/Email: I.R.Dodkins@swansea.ac.uk
Contents Executive Summary 1 1. Introduction 2 Definition 2 1.1 History 2 1.2 Range of applications 3 Domestic wastewater treatment 3 Metals treatment 3 Agricultural waste 3 Habitats 3 Tourism 4 2. Processes 4 2.1 Phosphorous removal 4 Settling and peat accretion 4 Soil adsorption 5 Precipitation of P 5 Plant uptake 5 Microbial and Algal uptake 5 Fish uptake 5 Problems with phosphorous removal 6 2.2 Nitrogen removal 8 2.2.1 Nitrogen removal in aerobic water 8 Ammonification (mineralisation) 8 Ammonia volatilisation 8 Nitrification 8 2.2.1 Nitrogen removal in anoxic water 9 Denitrification 9 Anaerobic Ammonia Oxidation (ANAMMOX) 10 Plant uptake 11 Problems with nitrogen removal 12 2.3 Oxygen 12 Factors influencing oxygen concentrations 12 Temperature affects 12 Plants 13 2.4 Redox potential 14
2.5 BOD, Suspended Solids and Carbon 15 Biological Oxygen Demand 15 Carbon 15 2.6 Metal removal 16 2.7 pH 18 2.8 Harvesting of Floating Island Plants 18 Practicalities of harvesting 19 Storage of nutrients in plants 19 Storage of metals in plants 20 Relative importance of different processes 20 3. Treatment Efficiency 21 Overview 28 3.1 Seasonal Variation 30 Rainfall 30 Shading and Temperature 31 4. Design Considerations 32 4.1 Island Cover 32 4.2 Optimising for N removal 32 Aerobic and Anaerobic basins 32 Recycling 33 4.3 Plants 33 Plant dimensions 33 Plant establishment 34 Buoyancy of islands 35 4.4 Activated Carbon 35 5. Conclusions and Recommendations 37 Recommendations for domestic effluent treatment 38 Use of Activated Carbon 38 6. References 39
Literature Review:
Floating Treatment Wetlands (FTWs) in Water Treatment: Treatment
efficiency and potential benefits of activated carbon.
Dr. Ian Dodkins; Anouska Mendzil; Leela O’Dea
Executive Summary
Floating Treatment Wetlands (FTWs) have many benefits over Free Water Surface (FWS)
wetlands:
1. Plant roots assisting in filtering and settling processes for sediment bound P and
metals
2. Plant roots acting as a large surface area for micro-organism activity in:
decomposition, nitrification, and denitrification (removal of BOD and N).
3. Mild acidification of water due to release of humic acids; and a C input from
senescent vegetation, assisting denitrification.
4. They can adjust to varying water levels
5. A higher retention time is possible as they can be made deeper without submerging
the vegetation
Percentage removal of nutrients and metals from effluent is around 20-40% higher in FTWs
than in conventional FWS ponds. Removal efficiency, particularly of nitrogen, can be further
increased with tighter control on the water chemistry (aeration; adding CaCO3; adding a
carbon source). 20% coverage of islands is optimal for aerobic basins. 100% cover is optimal
for anaerobic basins or aerobic basins where there is artificial aeration. The design the FTW
and the control of basin water chemistry is essential for optimising treatment efficiencies.
The passive use of activated carbon within layers of floating islands is unlikely to be cost
effective.
Page 1 of 44
Introduction
Definition
Floating Treatment Wetlands (FTWs) comprise of wetland basins or cells, on which there are
artificial mats containing emergent plants (Figure 1). This is not to be confused with
treatment using floating leaved plants such as Eichhornia crassipes (Water Hyacinth), Pistia
stratiotes (water lettuce), Lemna spp. (duck weed) or Azolla spp. (water fern) e.g. Reddy &
Smith (1987); Kivaisi (2001), or where natural floating islands have established. Floating
Treatment Wetlands are also referred to as Constructed Floating Wetlands (CFWs) or
Floating Mat Constructed Wetlands, but we will use FTW throughout the review. Floating
Islands (FIs) will be used to refer only to the islands within the treatment system. ‘Effluent’
refers to the water being treated at any stage within the wetland and ‘inflow’ refers to
effluent entering the wetland, and ‘outflow’ as effluent leaving the wetland. Comparison will
regularly be made between FTWs and other wetlands. Where ‘conventional wetlands’ is
referred to, this means other treatment wetlands in general. Basins where there is open
water but no islands, are known as Free Water Surface (FWS) wetlands.
The core of this review assesses process, performance and design of FTWs and includes a
section on the potential for incorporating activated carbon into FTWs.
Figure 1. A Floating Treatment Wetland (FTW). Emergent plants are grow within a floating artificially constructed
material. The roots are directly in contact with the effluent and can intercept suspended particles. The roots also
provide a high surface area for microbiological activity. Image: Headley and Tanner (2006).
1.1 History
Floating islands are a natural occurrence, and can be found where emergent aquatic plants
have broken from the land, sometimes developing in highly nutrient rich or sulphurous pools
(Duzer, 2004). Floating leaved plants for treatment date back to the 11th Century, when the
floating Azolla fern was used by Chinese and Vietnamese farmers to extract dissolved
nutrients from wetlands and rice paddies, after which it was dried and applied as soil
fertiliser (Whitton & Potts, 2002). The use of Water Hyacinth (Eichhornia crassipes) to
remove nutrients also developed in South East Asia, and both have been used for centuries
for water treatment within this region (Whitton & Potts, 2002). E. crassipes was suggested
for use in the early 20th Century in both Auckland (Australia) and Yorkshire (UK) (Dymond, no
date), and then in 1975 NASA used it to treat a sewage lagoon in the USA (Wolverton &
Mcdonald, 1978).
Constructed floating islands were first developed in Japan in the 1990s, with Cana generalis
being grown in floating beds to absorb nutrients from fish ponds and treatment basins (Wu
Page 2 of 44
et al., 2000). Twenty percent coverage of soilless artificial floating islands, again using C.
generalis, was later recommended to improve water quality in China (Bing & Chen, 2001).
Floating mats have also developed unintentionally in many open water treatment systems.
Sometimes detrimental effects were observed, such as in Florida, where mats which had
grown to 50% coverage moved with the wind across a shallow basin; scraping the bottom
and disturbing sediments, resulting in increased outflow turbidity and phosphorus release
(Kadlec & Wallace, 2009).
1.2 Range of applications
FTWs can be specifically designed or they can be installed in currently operating open water
wetlands i.e. retrofitting. Potentially FTWs can be used with the same waste streams as
conventional wetland systems. Some examples from the literature are:
Domestic wastewater treatment
Conventional vegetated wetlands have often been advocated for wastewater polishing,
rather than heavy nutrient loads, since they can become clogged and plants are very good at
removing low concentrations of nutrients. However, given that there is primary
sedimentation, FTWs can potentially deal with larger nutrient loadings since they have
higher P and N removal capacity compared to conventional wetland treatment systems. The
exposed roots aid sediment deposition, thus reducing turbidity, and there is greater surface
area available for microbiologically mediated nitrification/denitrification reactions.
Treatment with floating islands has been done on domestic waste in a highly controlled
environment i.e. as a hydroponic system (Vaillant et al., 2003).
Metals treatment
Wet detention ponds are used as a Best Management Practice for stormwater run-off in the
USA (Chang et al., 2013). FTWs have thus become a popular choice as a retrofit for
stormwater run-off treatment in these ponds (Chimney et al., 2006; Headley & Tanner, 2006;
Tanner & Headley, 2008, 2011; Hwang & Lepage, 2011; Chang et al., 2012; Borne et al., 2013;
White & Cousins, 2013; Winston et al., 2013). FIs are beneficial since they can treat water
effectively even with the large fluctuations in water depth that occur during storms.
Strosnider & Nairn (2010) stated that FTWs are ideal for acid mine drainage, particularly if
anaerobic conditions are maintained using high island cover. The resulting anaerobic
conditions and the decomposing plant material aids denitrification, making the water more
alkaline.
Agricultural waste
The enhanced nitrate removal rate of FTWs makes them appealing in reducing pollution
from agricultural run-off (Stewart et al., 2008; Yang et al., 2008) as well as for more
concentrated wastewaters, such as from swine effluent (Hubbard et al., 2004).
Habitats
FIs have sometimes been constructed specifically to create habitats e.g. to protect birds
from land-based predators (Hancock, 2000), including a huge floating island of 3 700 m2 in
Sheepy lake (California) as a habitat for nesting Caspian Terns (Patterson, 2012). Only islands
designed for effluent treatment will be covered in this review, however FIs do provide
habitats as a secondary function. Emergent grasses can attract waterfowl and terrestrial
birds because of the seeds, nesting material, nesting cover and available water. Fish have
been introduced into some open water wetlands, however those that feed or nest on the
Page 3 of 44
bottom have been found to disturb sediments, increasing suspended solids (Kadlec &
Wallace, 2009, p.779).
Although usually not problematic, there have been incidences where large bird communities
have contaminated open water treatment wetlands with faeces (Orosz-Coghlan et al., 2006),
or have disturbed sediments, increasing turbidity (Knowlton et al., 2002). Geese herbivory
can devastate the establishment of wetland plants, especially if planted during the spring or
autumn migratory period (Kadlec & Wallace, 2009). However, a benefit of floating islands is
that other herbivores e.g. rabbits, cannot usually access the islands. Mosquitoes may also be
a problem with an open water system, particularly where monotypic vegetation such as
cattail, bulrush and common reed restrict predator access (Knight et al., 2004). However,
removing leaf litter, and ensuring that water depth is greater than 40cm (Sinclair et al., 2000)
can reduce the problem.
Tourism
Treatment wetlands have been effectively marketed for tourism, especially those which
provide good natural habitats for birds (Kadlec & Wallace, 2009). If the FTW is operated for
tourism the design and operation is likely to have to include walkways, bird viewing areas
and education centres. There may also be conflicting aims for depth regulation between
habitat provision and treatment.
2. Processes
FTWs, as with other wetland treatment systems, remove pollutants by four main processes
(in order of importance): physical; biogeochemical; microbial and plants. These processes
are similar in conventional wetlands, so much of the details provided here comes from that
research. However, the larger surface area created by plant roots in FTWs tends to increase
sedimentation (by filtering), microbiological decomposition, nitrification and denitrification,
and also alter the water chemistry i.e. pH and dissolved oxygen (DO) concentrations.
Processes will be discussed relative to the effluent constituents being removed.
2.1 Phosphorous removal
Phosphorous within wetland effluents is usually as dissolved orthophosphate (PO43-), or
organic phosphorus (Masters, 2012). The scarcity of P in natural environments results in
efficient nutrient cycling within ecological systems (Kadlec & Wallace, 2009), thus there are
few permanent routes for removal of P within treatment wetlands (Figure 1). The major
mechanisms for P removal are accretion in peat/soil and soil adsorption.
Settling and peat accretion
Settling is the main process by which phosphorous bound sediments and BOD are removed
from the water column (Kadlec & Wallace, 2009). Settling is a physical process whereby
phosphate bound in particles sink to the bottom. Settling is increased in FTWs both by the
roots (Masters, 2012) which filter the particles from the water column to later slough off to
settle on the bottom, and by reducing currents and circulation caused by surface wind
disturbance or water movements (e.g. from pumps) (Headley & Tanner, 2006; Chang et al.,
2013). The reduction in movement is essential for preventing resuspension of sediment
bound phosphorous into the water column, however, this reduction in currents also
contributes to the risk that the basin will become anoxic (Van de Moortel et al., 2010). P
retention within different conventional wetlands ranges from 40-60%, around 45 to 75
g/m2/yr (Vymazal, 2007), most of this being due to settling (and associated processes such as
accretion and soil adsorption). P removal from FTWs is usually higher due to the additional
filtering properties of the roots, reaching 81% (White & Cousins, 2013).
Page 4 of 44
Soil adsorption
Phosphorus is retained in the soils by binding to the soil surface. Soils with high clay content
have high P adsorption capacity, which increases with lower pHs. Organic soils also adsorb P,
with the adsorption capacity dependent on mineral components (Rhue & Harris, 1999). Al
and Fe fix phosphorus in acidic soils, whilst Ca and Mg fix it in alkaline soils (Kadlec & Knight,
1996). This adsorption process is reversible, with an equilibrium between the bound P and
the dissolved P in the soil porewater. The soil minerals and binding sites result in a
‘phosphate buffering capacity’ which determines where this equilibrium exists (Barrow,
1983). This has important implications for P removal, since reducing inflow P can cause P
desorption from the sediments, actually producing a higher P outflow than inflow (Belmont
et al., 2009).
Precipitation of P
P adsorption occurs in aerobic waters, but as conditions become anoxic (reducing conditions)
metals within the soil change valency, becoming soluble. This causes the release of
phosphorus as a co-precipitate (precipitating due to the action of a true precipitate) from
the soil (Kadlec & Wallace, 2009). In very low oxygen conditions, where the soils are
anaerobic (Eh < -200 mV) sulphate reduction occurs (Figure 4). This creates free sulphide
which preferentially binds with Fe (as iron sulphide) preventing iron mineralisation of P.
Thus, anaerobic conditions promote the release of P back into the water column (Kadlec &
Wallace, 2009).
Plant uptake
Plant uptake of P reaches only around 6% (Masters, 2012). If a FTW has a P removal up to 81%
(White & Cousins, 2013), this means around 75% is removed predominantly by settling or
storage in other sinks. Much of the P in plant uptake is also difficult to remove permanently
from the system by harvesting because it is stored in the roots, or it re-enters the system as
litter (see Section 2.8 Harvesting). Vymazal (2007) considers that harvesting of conventional
wetlands is only useful in low P effluents (e.g. polishing) with around 10-20 g P/m2/yr, where
uptake is not limited by growth rate. FTWs may be able to absorb more P, due to their roots
being suspended directly in the effluent, and plant roots are more accessible for harvesting,
but dredging is still likely to be the most effective method of permanent removal.
Microbial and Algal uptake
Bacteria and algae are important in P cycling within the soils, rhizosphere and water column
(Vymazal, 2007). P uptake by microbes in conventional wetlands is very fast, but they store
very little (Vymazal, 2007). Thus, having higher surface area and consequently higher
microbial mass, microbes in a FTW are likely to be a larger sink of P than in conventional
treatment wetlands, however nutrient cycling is likely to result in little net removal, except
through sedimentation of dead organic microbial matter.
Fish Uptake
In South East Asia it is common to use fish for nutrient recovery in ponds receiving human
effluent (Cairncross & Feachem, 1993). Fish eat periphyton (such as algae, cyanobacteria,
heterotrophic microbes, and detritus) (Azim et al., 2005) as well as fungi, protozoa,
phytoplankton, zooplankton, invertebrates and invertebrate larvae, and some species are
piscivorous. In treatment wetlands fish are usually chosen for their adaptation to low oxygen
levels, for example Gambusia affinis (mosquito fish) in warm temperate to tropical
conditions, and Notrophus fundulus (black-stripped top minnow) or Umbra limi (central
mudminnow) in temperate climates with over 77 different fish species being used in North
American treatment wetlands (Kadlec & Wallace, 2009). Sometimes Oreochromis spp.
Page 5 of 44(Tilapia) and Bass have colonised previously unpopulated treatment wetlands (Kadlec &
Wallace, 2009).
Li & Li (2009) examined nutrient removal from aquaculture effluent using floating islands (17%
cover) planted with the aquatic vegetable Ipomoea aquatic. There was artificial aeration and
it was populated with Aristichthys nobilis (silver carp), Siniperca chuatsi (mandarin fish;
carnivorous) and Carassius auratus gibelio (crucian carp). Around 34% of TN and 18% of TP
was removed from the system, and of this around a third (34%) of removed TP and TN was
removed by fish. This was around the same that was removed by sedimentation.
Kania (2014, unpublished) suggests that FTW facilitate the sustainable growth of fish and
demonstrates that FTW significantly increase fish biomass that can be harvested from the
waterway. Fish harvesting enables P removal from the effluent with fish being made into
meal which can be used for pork or poultry farming or in pet food. There must be no toxins
or toxic metal contaminants in the effluent, especially contaminants that may bioaccumulate.
Also, if it is to be sold for human consumption the fish need to be cooked well since there is
the potential for contamination by pathogens, particularly the tapeworm Clonorchis sinensis
(Cairncross & Feachem, 1993).
Fish can disturb bottom sediments, releasing P, particularly those that feed or nest on the
bottom e.g. Cyprinus carpio (Carp) (Kadlec & Wallace, 2009, p.696).
Problems with phosphorous removal
Generally, wetland treatment only produces temporary storage of P, in contrast to N and C
which can be released as gases through microbiological degradation (N2 and CO2). Indeed,
Yousefi and Mohseni-Bandpei (2010) stated that P can be considered as a conserved entity.
Most P is stored in sinks such as sediments (95%; Masters, 2010), plants, microbes and algae,
but this P is recycled. These sinks give an initial period of apparent P removal. However,
once the wetland is established, nutrient cycling results in similar outflow P levels to inflow.
Even regular harvesting of plants only removes around 6% (Masters, 2012) of P inflow, if
both the roots and shoots are harvested. Thus, Kavanagh & Keller (2007) concluded that at
least 90% of P eventually passes through a wetland system and is released in the effluent.
Some wetland treatment systems can even export more P than they receive, such as a
stormwater wetland in North Carolina which had median removal efficiencies of – 95% to
70%; at times exporting twice as much P as it was receiving (Line et al., 2008). This can occur
both due to physical disturbance of the sediments releasing P, the re-release of P from
biodegradation of organics (Sundaravadivel & Vigneswaran, 2001), or anoxia which can also
result in the sudden release of P as a co-precipitate (Maine et al., 2005).
Sudden P releases into the water column can potentially have other detrimental effects.
Since P is usually the limiting factor for biological activity in freshwaters (Schindler et al.,
2008), a large P release can result in nitrogen becoming limiting. This promotes the growth
of Cyanobacteria blooms which as well as producing harmful toxins, also extract N from the
atmosphere (Conley et al., 2009).
Masters (2012) is thus emphatic that dredging is important for long term removal of
phosphorus from a FTW. Kadlec and Wallace (2009) detail projected working life of different
types of wetlands with different soils, ranging from around 10 to 170 years, but dredging
around every 10 years (Masters, 2010) would be ideal for sustained P removal with most
effluents.
Page 6 of 44A minor route of P removal is phosphine (PH3). It is usually found in very low amounts (e.g.
47 ng/m3 of water in marshes),, mostly bound to sediments but with around 10% of this
dissolved in the water (Hana et al., 2010).
2010) However, it can be released from highly anaerobic
wetlands (Eh < - 200mV) (Gassmann & Glindemann, 1993) as phosphine gas. Devai and
Delaune (1995) calculated a gaseous release rate of 1.7 g P/m2/yr from a bulrush wetland
treatment system.
Thus, treatment wetlands have various sinks (algae, plants, microbes, soils) which vary in
their capacity to absorb P from the effluent based on conditions such as available surface
area, soil type, pH and redox potential.
potential FTWs limit the resuspension
esuspension of particulates since the
islands reduce water movement within the wetland and the roots filter out particulat
particulates
(Borne et al., 2013),, thus increasing P sedimentation. However, dredging is essential to long
term functioning
ctioning of a FTW for P removal, and regular harvesting can be useful at low P
loadings (Figure 1).
Harvesting
(before autumn as P
redirected towards
roots in autumn)
Dredging
Release of litter
Algae/ (autumn)
phyotplankton Plant uptake
uptake Leaching of
(mostly AEROBIC
nutrients
in spring)
PO43- PO43- PO43- (Eh >300 mV)
mV
0.06-75
P
Soil adsorption
Peat/
(depends on Al, Fe, Ca, Mg) P soil accretion P
Permanent
storage
0.5
ANOXIC
(Eh -100
100 to 300 mV)
mV
P
Dissolution and Precipitation with
Fe, Al, Ca, Mg
Figure 2. Summary of phosphorus
hosphorus processes in aerobic and anoxic wetlands.
wetlands
Soil/peat accretion and soil adsorption is the major process and major (95%) sink. However, sorption of P into the
P/m2/yr that
soil is reversible. Without harvesting or dredging P removal eventually stops. Numbers in bold are g P
may be removed or added during the processes;
process italics indicate
dicate the name of the process, with specific conditions
required in brackets.
Page 7 of 442.2 Nitrogen removal
Nitrogen is the principal target for treatment in many wetlands. Effluents contain organic
nitrogen compounds, which break down principally to ammonia, from which nitrite and
nitrate can form through a microbiological nitrification process. Different micro-organisms
within anoxic zones can denitrify this nitrate to permanently release N2 gas from the basin.
Agricultural wastes may already have high concentrations of nitrate nitrogen as they enter
the wetland. Conversion between different forms of N depends on many factors, including
DO, available carbon and pH.
2.2.1 Nitrogen removal in aerobic water
Ammonification (mineralisation)
Dead and decaying organic material is broken down in to ammonia by microbes, either
utilising the energy released or absorbing the ammonia for use as microbial biomass.
Ammonification increases with temperature, being optimal at 40-60 °C, and with organic
compound availability (especially when they have low C/N ratios) (Reddy & Patrick, 1984).
Optimum pH is between 6.5 and 8.5 (Vymazal, 2007). Ammonification usually takes place
under aerobic conditions (oxidative deamination).
Equation 1: Break down of organic N (example with amino acid) to ammonia
RCH(NH2)COOH + H20 → NH3 + CO2
Ammonification rates can vary greatly e.g. between 0.004 and 0.53 g N/m2/d (Reddy &
D’Angelo, 1997; Tanner et al., 2002). The root zone of a FTW is likely to be a good location
for ammonification.
Ammonia volatilisation
Ammonia exists in an equilibrium between its dissolved ammonium form (NH4+) and its
gaseous form (NH3); Equation 2. Below pH 8.0 ammonia loss as gas is negligible (Reddy &
Patrick, 1984). At a pH around 9.3 losses due to volatilisation can become significant
(Vymazal, 2007). N removal rates due to ammonia volatilisation have been measured at 2.2
g N/m2/d in wetlands (Stowell et al., 1981).
Equation 2: Conversion of dissolved ammonium to ammonia gas
NH4+ + OH- NH3 + H2O
Algal photosynthesis often elevates pH values during the day (Vymazal, 2007), thus
increasing ammonia volatilisation. However, FTWs may inhibit this due to (i) islands shading
algae and reducing the area of the air-water interface, and (ii) plants releasing humic acid,
which reduces the pH (Van de Moortel et al., 2010).
Nitrification
Within aerobic water micro-organisms convert ammonium to nitrate in a process called
nitrification. Directly adjacent to plant roots there is an aerobic zone (Reddy et al., 1989),
which means that FTW are likely to have elevated denitrification rates due to the availability
of root surface area.
Kadlec & Wallace (2009; p.280) note that nitrification in wetlands is quite different from
nitrification in conventional Waste Water Treatment Works (WWTWs). Whilst nitrification is
Page 8 of 44commonly considered a two step process in conventional WWTWs, in natural wetlands it is
now believed to have three stages (Bothe et al., 2000); Equation 3.
Equation 3: The three stage nitrification process, converting ammonium to nitrite, then nitrate.
Nitritation (2 stages)
Nitrosomonas
NH3 + O2 + 2H + 2e- ----------------------------→ NH2OH + H20
Nitrosomonas
NH2OH + H20 ----------------------------→ NO2- + 5H + 4e-
Nitrification (1 stage) Nitrospira or Nitrobacter
2NO2- + O2 ----------------------------→ 2NO3-
Due to the different processes less oxygen and alkalinity is consumed in wetlands during
nitrification than in conventional WWTWs (Kadlec & Wallace, 2009). Nitrospira is also much
more prominent as a nitrifier than Nitrobacter in wetlands (Austin et al., 2003).
Nitrification is influenced by temperature (optimum 25-35 °C), pH (optimum 6.6-8), alkalinity,
microbial populations present, DO and ammonium concentrations (Vymazal, 1995). Below
4 °C nitrifying bacteria Nitrosomonas and Nitrobacter do not grow (Paul & Clark, 1996).
Kadlec & Wallace (2009; p.280) note, unlike WWTWs, there is little evidence that a low C/N
ratio in wetland effluents improves nitrification rates.
In wetlands, for every g of ammonium oxidised to nitrate 2.28 g of oxygen and 7.1 g of
alkalinity as calcium carbonate are consumed (Kadlec & Wallace, 2009; p.279) i.e.
nitrification requires aerobic conditions and will consume alkalinity and oxygen, becoming
increasingly acidic and anaerobic. Wetlands have nitrification rates of 0.01 to 2.15 g N/m2/d
(mean of 0.048) (Reddy & D’Angelo, 1997; Tanner et al., 2002), though this may be much
higher for FTWs due to the large root surface area within the aerobic zone.
Low oxygen conditions can result in nitrite (NO2-) being produced instead of completing the
process toward nitrate (Bernet et al., 2001). The consequence of this is that in a later
denitrification stage, some of the nitrite is converted into nitrous oxide (N2O), a potent
greenhouse gas. Sufficient oxygenation in nitrification basins is therefore recommended.
2.2.1 Nitrogen removal in anoxic water
Denitrification
Denitrification is the microbiologically mediated conversion of nitrate into nitrogen gas,
which is then released from the wetland into the atmosphere. A carbon source is required
for denitrification. The equation can be written in many ways, depending on the source
assumed (Equation 4).
Equation 4: Denitrification of methanol, producing nitrogen gas and alkalinity
6NO3- + 5CH3OH-------------→ 3N2 + 5CO2 + 7H20 + 6OH-
In many ways denitrification is the converse of nitrification, making the water more alkaline
and requiring anoxic or anaerobic conditions. Microorganisms denitrify because in the
absence of dissolved oxygen for reduction, they reduce nitrate. Although methanol is used
Page 9 of 44for illustration here as a source of carbon, usually it is large organic molecules. It is
calculated that per g of NO3- around 3.02 g of organic matter (or 2.3g of BOD) is consumed,
and around 3g of alkalinity as CaCO3 is produced (Kadlec & Wallace, 2009).
The optimum pH is 6 to 8 (Paul & Clark, 1996) being negligible below pH4 (Vymazal, 2007).
Denitrification is very slow below 5 °C, but increases with temperature up to 60 or 75 °C,
then decrease rapidly (Paul & Clark, 1996). More nitrate can speed up the process, but the
limiting factor in denitrification is often the carbon supply (Kadlec & Wallace, 2009),
especially if BOD has settled out in previous treatment basins. A C/N ratio of 5:1 is suggested
to ensure carbon does not become limiting (Baker, 1998) although this may be an
overestimate if much of the C is labile (Kadlec & Wallace, 2009). Lower pHs can assist with
breaking down lignin in cell walls, increasing the litter quality for denitrification processes
(Ding et al., 2012).
Often an anaerobic denitrification basin is placed after an aerobic nitrification basin. This
enables all the ammonium to be converted to nitrate prior to denitrification, thus
maximising total N removal. However, even in a well oxygenated basin there are areas of
low mixing, and deeper waters and sediments, where oxygen levels are low enough to
produce denitrification (Figure 3, Figure 4), and in anoxic basins nitrification can occur on the
surface of roots where the plants have transported oxygen (Kadlec & Wallace, 2009, p.281).
Thus both nitrification and denitrification processes can be achieved within a single basin,
though controlling the treatment efficiency may be more difficult.
Floating islands can aid denitrification by producing anoxic conditions through the restriction
of oxygen diffusion into the water column. Also, roots and plant litter, as well as coconut coir
on islands (Baquerizo et al., 2002),can act as sorption sites, with biofilms developing which
increase denitrification rates and thus NO3 removal rates (Vymazal, 2007). Denitrification
releases are about 0.003 to 1.02 g N/m2/d in wetlands (Vymazal, 2007), though this could be
higher in FTWs due to more biofilm area and more sorption sites.
Anaerobic Ammonia Oxidation (ANAMMOX)
The bacteria involved in this process were only discovered in 1999. Planctomycetes
Nitrosomonas eutropha utilises ammonium ions and nitrite (from nitrification of ammonium)
to produce nitrogen gas. This can be represented as in Equation 5.
Equation 5. The ANAMMOX process
Formation of nitrite
2NH4+ + 3O2 -------------→ 2NO2- + 4H+ + 2H2O
ANAMMOX
NH4+ + NO2- -------------→ N2 + 2H2O
This denitrification process uses less than half the oxygen (1.94g O per gram of NH4+) of the
standard denitrification process, and requires no carbon substrate (Kadlec & Wallace, 2009).
ANAMMOX processes occur in many types of wetlands when there is severely restricted
oxygen. Bishay & Kadlec (2005) found that in a Free Water Surface wetlands there were
more ammonia losses than could be accounted for by the oxygen consumed under normal
dentification. There was also a lot of nitrite present in this wetland, and very little carbon,
suggesting that these conditions were conducive to the ANAMMOX reaction.
Page 10 of 44Plant uptake
Nitrogen uptake by plants in conventional wetland treatment is low (up to 6-8%) 8%) compared
to microbial denitrification (up
up to 61-63%) (Metheson et al., 2002). Vymazal (2007)
2007) estimates
that for conventional wetland systems plant harvesting is useful for N removal if i loading is
2
only around 100-200 g N/m /yr. /yr. If N removal is a priority, designing and operating the basins
to maximise nitrification/denitrification by microorganisms
microorganisms is probably more cost effective.
N is predominantly taken up by plants in the form of ammonia, but also as nitrate.. Much
M of
this is returned to the system when tissues senesce (Kadlec & Wallace, 2009).
0.6 - 88 Harvesting
(prior to autumn)
NH3
Volatilisation
2.2 (pH > 7.5)
Nitrification
(optimal 30-40°C;
pH 6.6-8.0)
NH4+ NO3-
0.001 – 2.15 AEROBIC
NH3 NO3- (Eh >300 mV)
Ammonification
CxOyNz NH3 (optimal at 40-60°C;
Amino pH 6.5-8.5) NH3
Organic
Ammonia adsorption acids
nitrogen burial
0.004 – 0.53
N2 N2 N2
Fixation by
cyanobacteria 0.003 - Denitrification
0.03 - 46.2 ANAMOX (Eh
Eh 350 to 100mV;
(Eh -200 to -260mV
260mV)
denitrification 1.02 pH 6-8;)
NH4+ NO2- NO3-
NH3
ANOXIC
(Eh -100
- to 300 mV)
Figure 2. Summary of nitrogen processes in aerobic and anoxic wetlands. Primary settlement of effluent is
assumed prior to entering the wetland. Numbers in bold are g N/m2/yr that may be removed or added during the
process; italics indicates the process,
cess, with specific conditions required in brackets. The striped blue arrow
indicates nitrogen fixation that would not normally occur unless the anoxic pond becomes anaerobic (Eh < -200
mV).Permanent removal of N is only through ammonia volatilisation (minor), or), denitrification (including ANAMOX)
and harvesting. Organic nitrogen burial (associated with litter) and ammonia adsorption (associated with clay
soils) are relatively minor processes.
Page 11 of 44Problems with nitrogen removal
NH4 removal rates in conventional wetlands vary between 35 and 50% in Europe (Verhoeven
& Meuleman, 1999; Vymazal, 2002). FTWs have shown removal rates from -45% to 75% for
NH4 and between 36% and 40% for total nitrogen (Boutwell, 2002; DeBusk & Hunt, 2005;
Gonzalez et al., 2005).
Problems with nitrogen removal are associated with producing the correct microbiological
conditions; aerobic for nitrification and anoxic for denitrification, as well as ensuring
sufficient carbon supply for the later. These are discussed in the design section.
2.3 Oxygen
Factors influencing oxygen concentrations
Wetlands typically have slow flow, incomplete mixing, and rapidly decreasing oxygen profiles
with depth (Figure 3). Anoxic zones develop just below the substrate in shallower basins and
also in the lower regions of the water column in deeper basins (Kadlec & Wallace, 2009).
Oxygen can be rapidly depleted in wetlands due to microbiological activity, particularly with
nitrification and decomposition (Kadlec & Knight, 1996). FTWs exacerbate oxygen depletion
both due to high rates of microbiological activity (nitrification) and due to the islands
restricting diffusion of oxygen back in to the water i.e. reducing air-water contact area and
reducing wind disturbance (Van de Moortel et al., 2010). This makes FTWs particularly
susceptible to unwanted drops in DO, especially at high percentage cover of islands.
Figure 3. Vertical profiles of dissolved oxygen in various types of FWS (Free Water Surface) wetlands, Florida.
Data from 141 profiles collected over a 2½ year period. Data from Chimney et al. (2006), Figure from Kadlec &
Wallace (2009). FTWs are most readily compared to floating plant systems.
Temperature affects
Oxygen saturation of water varies with temperature: at 25 °C dissolved oxygen is 8.2 mg/l,
and at 5 °C it is 12.8 mg/l. However Kadlec and Wallace (2009) note that the poor mixing of
waters limits the dissolution of oxygen such that reaeration is very slow, even in open
Page 12 of 44wetlands. They estimate that it takes 2 to 4 days to reaerate an open wetland basin from 0
to 90% DO, with typical winds. This is likely to be even slower in FTWs.
Plants
Submerged photosynthesising plants and algae release in the range of 0.26 and 0.96 g/m2/d
of O2 during photosynthesis (Kadlec & Wallace, 2009, p138), oxygenating the water.
Emergent plants bring O2 to the roots, but O2 delivery usually matches respiration
requirements, so there is little net input into the water column (Brix & Schierup, 1990).
Studies by Tanner and Headley (2011) and White and Cousins (2013) both found a high level
of oxygen depletion in basins due to floating islands.
Tanner and Headley (2011) not only illustrated how oxygen depletion is higher in FTWs, but
also that oxygen depletion is higher when there are plants rather than mats with artificial
roots (sisal) (Table 1). This oxygen depletion is likely due to the higher rate of microbiological
activity associated with plant roots. Although the relationship between oxygen depletion
and root biomass was weak, there was little oxygen depletion due to the floating mat alone
and even the mat with artificial roots.
Table 1. Oxygen depletion (%DO) at subsurface and bottom of mesocosms after 7 days due to the effect of
Floating Islands, ordered from highest to lowest. Influent DO was 95%, floating island coverage was 50%. Root
biomass (dry weight) also shown. Adapted from Tanner and Headley (2011).
Root
subsurface bottom biomass
DO (%) DO (%) (g/m2)
Control (no floating mat, but equivalent shading) 87 85
Floating mat only 85 84
Mat + soil + artificial roots 85 84
Mat + soil media 80 79
Mat + soil + Juncus edgariae 68 66 299
Mat + soil + Schoenoplectus tabernaemontani 68 67 184
Mat + soil + Carex virgata 58 57 533
Mat + soil + Cyperus ustulatus 50 48 329
Van de Moortel et al. (2010) found redox potentials to be decreased due to floating islands:
at both 5cm and 60cm depths the FTW has much lower O2 than an open water basin: at
5cm redox is 68 (open water) cf. -25 (FTW); at 60cm redox is: -93 (open water) cf. -122 (FTW).
They did claim that roots can aerate island matting. However, there was little difference
between the mat redox potential (72 mV ± 478) and the redox potential 5cm below the
surface of an open water basin at (68 mV ±225).
A liability with FTWs is that during summer periods, due to high rates of microbiological
activity and insufficient O2 exchange with the atmosphere, the basin can become anaerobic,
causing sulphide toxicity which then kills the plant roots (Lamers et al., 2002) and
consequently reducing the effectiveness of treatment. Reduction in treatment efficiency due
to anoxia was found in several studies, but usually when the floating islands occupied 50% or
more of the surface water area (Van de Moortel et al., 2010; Borne et al., 2013).
Page 13 of 442.4 Redox potential
Oxidation is the loss of electrons during a reaction. This is usually through a substance
combining with oxygen, as it is energetically the most favourable oxidant. Redox potential is
the tendency of a system to oxidise substances i.e. in high redox potential water, incoming
organic substances will be rapidly oxidised (an oxidising environment) whereas in low redox
potential waters substances will be reduced (a reducing environment). An example of
reduction would be where hydrogen combines with carbon to produce methane.
Redox potential is strongly associated with the oxygenation of the water, but it is not
identical, since substances other than O2 can oxidise. Zonation usually occurs in a wetland
with oxygen being the oxidiser near the surface, then as DO decreases other substances
become oxidisers, with reactions releasing less energy with successively weaker oxidisers.
This is in the order O2, NO3-, MnO2, FeOOH, SO42- then CO2.
The decline in free oxygen reflects the redox potential (Eh), also known as the oxidation-
reduction potential (ORP), of the water i.e. the tendency of a chemical to acquire electrons,
measured as electric potential (mV). At Eh > 300mV (measured with a platinum electrode)
conditions are considered aerobic, at < -100 mV conditions are anaerobic, and between
these (near-zero Dissolved Oxygen) conditions are anoxic (Figure 4).
Redox Potential Reactions Zone
> +300 mV Oxygen reduction I Aerobic
---
+ 100 to +300 mV NO3- and Mn4+ reduction II
+100 to – 100 mV Fe3+ and Mn3+ reduction III Anoxic
-100 to -200 mV SO42- reduction IV
Anaerobic
< -200 mV CH4 formation V
Figure 4. Redox zonation in wetlands, based on Kadlec & Wallace (2009). This vertical zonation can be found in
deep lentic environments, particularly where there is high oxygen consumption e.g. my microorganisms.
At high redox potential phosphorus can form insoluble complexes with oxidised iron,
calcium and aluminium. Organic compounds which comprise most of the BOD are oxidised
using oxygen by bacteria, releasing carbon dioxide. At lower redox potentials organic
material does not decay quickly. The water is anoxic, with reducing conditions
predominating. Manganese and iron are both reduced (Equation 6)
Page 14 of 44Equation 6. Reduction of manganese and iron in anaerobic conditions
Mn4+ + 2e- → Mn2+ Fe3+ + e- → Fe2+
This reduction causes metals to precipitate out of the sediments back into the water column,
bringing P with them, as a co-precipitate (Van de Moortel et al., 2010).
Further decreases in oxygen (below -100mV) result in anaerobic conditions, whereby
sulphate is reduced to hydrogen sulphide, which although soluble, can be released as gas at
low pH (Kadlec & Wallace, 2009). Usually this reduction is undesirable in wetlands, except in
acid mine treatment.
Equation 7. Reduction of sulphate to hydrogen sulphide
SO2-4 + 2CH2O → H2S + SHCO3-
Eventually, at very low redox potential (below -200mV) CO2 , formate, or acetate, is reduced
to methane (CH4) by bacteria.
Equation 8. Reduction of carbon dioxide to methane.
4H2 + CO2 → CH4 + 2H20
2.5 BOD, Suspended Solids and Carbon
Biological Oxygen Demand
Biological Oxygen Demand (BOD) is a measure of oxygen consumption by microorganisms
due to the oxidation of organic matter; usually measured in the lab over 5 days (BOD5). BOD
of inflows are typically high, unless the treatment basin is being used just for polishing
previously treated wastes. BOD decreases rapidly (around 50% decrease within 6 hours) as it
passes through a wetland due to decomposition and settling of organic carbon, finally
reaching a non-zero plateau (Kadlec & Wallace, 2009). Even if the waters are not aerobic,
fermentation and sulphate reduction can remove carbon from the system.
Carbon
Most carbon entering a wetland is organic. Microbiological processes are the main method
for removing carbon, through the oxidation of organic compounds, releasing energy. In
aerobic waters, respiration takes place (Equation 9), releasing CO2 to the atmosphere. In
anaerobic zones there are four main processes which can take place: (i) fermentation
producing either lactic acid or ethanol (ii) methanogenesis producing gaseous methane (iii)
sulphate (SO42-) reduction producing carbon dioxide and hydrogen sulphide, and (iv)
denitrification, producing carbon dioxide and gaseous nitrogen.
Settling is also an important removal method (although the carbon is retained in the
sediments). In FTWs plants have been shown to remove around 5.9 g BOD/m2/day. The large
surface area provided by roots can produce a higher rate of microbial decomposition
(Brisson & Chazarenc, 2009), but roots also physically entrap particulates onto the biofilm
which then fall in clumps and settle out, providing a significant removal pathway for
suspended solids (Smith & Kalin, 2000; Headley & Tanner, 2006; Van de Moortel et al., 2010;
Borne et al., 2013). Settling is further encouraged by flow resistance through the roots and
flow reduction caused by wind shielding of the surface. Particulate carbon, and carbon
Page 15 of 44bound in litter, if it is not decomposed, accumulates in the sediments, particularly where
conditions are anaerobic (Kadlec & Wallace, 2009).
Equation 9. Microbiological decomposition of organic compounds.
Respiration
C6H12O6 → C02 + H20
Fermentation
C6H12O6 → 2CHCHOHCOOH (lactic acid)
C6H12O6 → 2CH3CH2OH + CO2 (ethanol)
Methanogenesis
(acetate) CH3COO- + 4H2 → 2CH4 +H20 + OH-
Sulphate reduction
(lactate) 2CH3CHOHCOO- + SO42- + H+ → 2CO2 + 2H2O + HS- + 2CH3COO- (acetate)
Denitrification (see Equation 4)
Unlike submerged plants, which obtain carbon from the water, carbon uptake by emergent
plants is from atmospheric CO2. Plants thus bring carbon into the system through
photosynthesis and the deposition of organic matter. However, the net effect of plants in
wetlands is to reduce BOD due to plant respiration, increased settling, and increased
decomposition processes (Masters, 2012). Also, where there is carbon limitation in anoxic or
anaerobic basins, the C provided by the deposition of litter can be important in increasing
denitrification rates (see Section 2.2.1).
Settling of BOD is also affected by basin depth, residence time and water movement (Kadlec
& Wallace, 2009). Theoretically higher temperatures should increase microbial
decomposition rates. Bacteria have limited activity below 5°C, but in conventional wetlands
there is no significant temperature dependence above this (Akratos & Tsihrintzis, 2007;
Kadlec & Wallace, 2009). This may be due to limitations in oxygen transfer rates or
restricting factors in one or more of the many C processes (Kadlec & Wallace, 2009).
In anoxic (reducing) conditions, the presence of sulphate contributes to the removal of
organic matter (BOD/COD) by acting as a coagulant and thus increasing settling rates (Huang
2005).
2.6 Metal removal
Metal removal from wetlands is predominantly through forming complexes with organic
matter, and through being coated in iron or manganese oxyhydroxides (Kadlec & Wallace,
2009). This either occurs in the sediments, or they settle out into the sediments. Under
anoxic conditions Cu, Zn, Pb, Ni and Ca form insoluble metal sulphides which will settle out.
Even in aerobic basins, decomposition of organic matter usually means there is an anoxic
layer just below the surface oxic layer (≈1cm) in which these metal sulphides can form.
Page 16 of 44Predicting metal removal from wetlands can be very difficult, depending on the structure of
the sediments and many factors of the water chemistry, with models regularly being wrong
by orders of magnitude (Kadlec & Wallace, 2009). Factors that affect metal removal include
the Cation Exchange Capacity (CEC) of the sediments, pH (circumneutral usually being
optimum), redox potential, the availability of sulphur for the formation of metal sulphides,
and the formation of iron and manganese oxyhydroxides (which allow co-precipitation)
(Kadlec & Wallace, 2009). Organic soils with humic acids and phenolics increase the CEC and
thus adsorption of metals. Sedimentation of metals can result in long term storage,
depending on the availability of organics with which metals can complex, although metal
accumulation can eventually saturate the soil sink and result in biological toxicity (Kadlec &
Wallace, 2009). Thus (careful) dredging is required in the long term to permanently remove
metals and ensure the wetland continues to operate effectively.
Uptake by plants is much less important than that by sedimentation, and where metals are
taken up, they are mostly stored in the roots. Table 2 shows percentage removal of metals
by plants in a conventional wetland and how this is allocated in the roots and shoots.
Table 2. Percentage removal of metals by plants in a conventional treatment wetland and how this is allocated to
the roots and shoots (adapted from Nolte and Associates, 1998).
Metal Roots (%) Shoots (%) Total (%)
Ag 2.0 0.0 2.0
As 10.1 0.6 10.7
Cd 13.3 0.0 13.3
Cr 16.8 2.2 19.0
Cu 5.5 0.6 6.1
Hg 6.7 0.0 6.7
Ni 4.7 0.3 5.0
Pb 11.8 2.0 13.8
Zn 6.1 0.4 6.5
Despite plant uptake being low, FTWs have been shown to greatly increase metal removal
compared to unvegetated retention ponds. For example Borne et al. (2013) compared
treatment in a normal stormwater retention pond with one retrofitted with a floating island.
With concentrations of 0.0092 mg Cu/l and 0.035 mg Zn/l in the inflow, particulate Cu and
Zn removal was 19% and 40% (respectively) in the normal pond, and 50% and 65% with a
floating island. Tanner & Headley (2011) found the removal of dissolved Cu and Zn to be 5%
and 1% without a floating island, and 50% and 47% with an island. These authors believe
that the benefit of the floating island wasn’t principally due to plant uptake. Indeed, Tanner
& Headley (2011) found mean plant uptake rates were 0.059-0.114 mg Cu/m2/d and 1.2-3.3
mg Zn/m2/d, accounting for less than 4% of Cu removal and less than 10% of Zn removal.
This was a mesocosm experiment without bottom sediments and with predominantly
dissolved metals, so values of plant uptake were probably higher than they would be in a
normal FTW.
Tanner & Headley (2011) and Borne et al. (2013) considered that the improved performance
with floating islands was due mainly to: (i) interception by the plant roots, (ii) humic acid
release from the plants, which reduced alkaline waters to circumneutral pH (Van de Moortel
et al., 2010), improving metal complexation and therefore flocculation and settling (Mucha
et al., 2008) and (iii) The islands reducing the redox potential to the extent that insoluble
metal sulphides formed.
Page 17 of 44The exact mechanisms of metal removal depend on the specific metal. Most zinc within
effluents is in particulate form and is removed predominantly through settling, sorption to
organic sediments and chemical precipitation/co-precipitation (Kadlec & Wallace, 2009). It
can form precipitates with sulphur (ZnS) and carbonate from the water (ZnCO3) and it co-
precipitates with Fe, Mn, Al oxyhydroxides. However, ZnS does not readily precipitate in
neutral waters (Younger, 2000), only in more alkaline waters (>7.5). Also, for co-precipiration,
the other metals must be present in the effluent, and even then, Fe and Mn oxides are not
stable in anoxic waters (Knox et al., 2004). Warmer water temperatures are also correlated
with Zn removal, probably due to increased sorption rates (Borne et al., 2013). Aerobic
wetlands are expected to absorb about 0.04g Zn/m2/d (PIRAMID consortium, 2003). Similar
to Zn, Cu removal rates increase with temperature, however adsorption is better at more
neutral pH (Borne et al., 2013). They also concluded that reduced oxygen resulted in a high
production of Cu sulphide precipitates in basins with floating islands.
High loadings of effluent and insufficient adsorption capacity or saturation of the potential
sinks (organic carbon, metal hydroxides, high CEC soils) can result in decreasing treatment
capacity as well as increasing toxicity. Toxicity can be a biological problem, particularly in
open water treatment systems where birds, amphibians and freshwater invertebrates have
direct access to the basin (as opposed to subsurface flow systems) (Kadlec & Wallace, 2009).
Sorption capacity in studies listed by Kadlec and Wallace estimate between 20 and 780 years
operation of a wetland with metal loading. Careful dredging (avoiding resuspension) can be
applied to remove contaminated sludges/soils. In mixed wastewater effluents from WWTWs
it is likely that the necessity for P removal through regular dredging is higher than that from
metal accumulation.
2.7 pH
pH has a profound effect on the functioning of wetlands, as mentioned in previous sections.
Several studies have confirmed the effect of floating vegetated islands in reducing pH. In a
two year study by White and Cousins (2013) pH decreased from 8.6 to 6.2. After only 11
days Van de Moortel et al. (2010) found a significant pH decrease from 7.5 to 7.0 whilst the
control (without an island) stayed constant at around 7.5. Borne et al. (2013) found a
difference between the control (8.3) and the FTW (7.3), which aided Cu adsorption.
Interestingly Tanner and Headley (2011) didn’t notice a drop in pH in mesocosm tanks,
although they still found that treatment was enhanced with floating islands, attributing the
difference in to the release of bioactive compounds. The researchers who found differences
in pH generally agreed that humic compounds were released by the plants, reducing pH.
White and Cousins also acknowledged that alkalinity consumed during microbial nitrification
on the plant roots could also be a driving force behind dropping pH within aerobic basins.
2.8 Harvesting of Floating Island Plants
FTWs are a relatively new technology with few long term studies, and few details on plant
harvesting. The prime functions of plants in FTWs is (i) for their roots to intercept and filter
particulates, aiding sedimentation, (ii) to increase the rates of microbiological processes by
providing a high surface area on which microorganisms respire, nitrify or denitrify, and (iii) to
alter the physic-chemical and chemical environment i.e. increase microbiological processing
through the release of humic acids and through reducing DO exchange (acidity and lower
oxygen increasing denitrification) and carbon deposition (increasing denitrification).
Harvesting is therefore not essential to long term management of FTWs, and although it can
help with permanent removal of nutrients and metals, removal rates are typically low. For
example, in subsurface flow wetlands plants only removed 2-8% of total nitrogen (Tanner,
2001; Yousefi & Mohseni-Bandpei, 2010) and 3-12% of total phosphorous (Yousefi &
Page 18 of 44Mohseni-Bandpei, 2010), with microbes believed to be removing the rest of the N, and
settling removing the rest of the P. Even with total uptake for N and P estimated at around
6%, all of this is unlikely to be harvested as it is stored in both the roots and shoots, and
nutrients are returned back to the wetland through deposition of senescent material.
Practicalities of harvesting
Floating islands facilitate easy harvesting. Often larger islands are able to support the weight
of humans, and so cutting could be done directly on the island. Smaller islands can be pulled
towards the shore and even lifted out. In contrast with other wetlands where the vegetation
is rooted in the sediments, in FTWs both roots and shoots can be removed, and with little
disturbance to the sediments. Theoretically a replacement island could be installed
immediately, although this may not be cost effective. Also, removal of root mass is likely to
be more detrimental to treatment than the gains from permanent removal of the nutrients.
For example, FTWs typically increase N and P removal rates by around 20-40% (Table 14),
whereas P and N removal by harvesting the whole plant is at the most 6%.
Storage of nutrients in plants
The start of the growing season, in early spring and prior to maximum growth rate, is the
time of highest P uptake. However, prior to autumn senescence, much of the P is relocated
to the root stock for the following year (Vymazal, 2007). Thus, if removal of P is a priority,
harvest timing and frequency is extremely important, with a recommendation that it is done
not only prior to senescence, but also during the peak growth period. The P lost in the
senescent material re-enters the basin system very rapidly; up to 30% lost through leaching
within the first few days of decomposition (Vymazal, 2007).
Although shoot biomass tends to be larger than root biomass (see plants in design section),
there is generally more N, P and K stored in the roots than in the shoots, especially when
autumn approaches (White & Cousins, 2013; Winston et al., 2013) e.g. Figure 4.
Figure 4. Nitrogen and phosphorus in roots and shoots of Canna flaccida and Juncus effusus after one summer of
2
growth (harvested 18 September 2008). Nutrients are per m of floating island. Three replicates per bar, with
standard error indicated. From White & Cousins (2013).
Page 19 of 44Storage of metals in plants
Storage of metals tends to show either an even distribution between roots and shoots (e.g.
Cu) or predominant storage in the roots (e.g. Zn) (Tanner & Headley, 2011). Table 3 shows
uptake of copper and zinc in roots and shoots over a 7 day trial.
Table 3. Uptake of copper and zinc in roots and shoots of four different plant species over 7 days in a FTW,
2
measured as µg/m /d. Adapted from: Tanner and Headley (2011).
Cu Zn
Plant species roots shoots roots shoots
Cyprus ustulatus 54 61 3027 282
Carex virgata 54 89 1228 934
Juncus edgariae 38 41 1703 760
Schoenoplectus tabernaemontani 36 24 881 320
Relative importance of different processes
Restricting flow and intercepting particulates on roots is one of the prime benefits of FTWs,
consistently removing more BOD and P than open water treatment ponds. However, using
synthetic root structure (sisal) Borne et al. (2013) showed that the physical structure alone
does not account for most of the benefits of FTWs; water chemistry changes, and to a much
lesser extent plant uptake, assist with improving treatment.
Page 20 of 443. Treatment Efficiency
Treatment efficiency obtained within a FTW is highly dependent on appropriate design and
proper operation, as well as the characteristics of the inflow and the objectives of the
treatment. At one end of the scale are FTWs designed for aerobic treatment with mixing or
air bubbled into the system, often with low % island coverage and addition of calcium
carbonate to aid nitrification reactions. These basins are predominantly to remove
ammonium. At the other end of the scale are anaerobic basins with up to 100% island
coverage, with addition of carbon in the form of e.g. molasses, to supply the denitrification
process. Thus, in aerobic basins, ammonium removal may be high whereas nitrate is
produced and may exceed inflow nitrate concentrations. In the latter, denitrification
reactions remove nitrate, but ammonium may not be nitrified, resulting in NH4+ increasing
(due to organic carbon decomposition) such that outflow exceeds inflow. Sometimes
floating islands achieve very high rates of removal because of a tightly controlled DO, pH and
carbon supply in a hydroponic system. The concentrations of pollutants also affects the
removal rate, with higher inflow concentrations often resulting in higher removal rates.
There can be different flow regimes, such as plug flow, where a quantity of effluent is kept in
the basin for around 3-7 days, continuous flow, or sporadic flow (such as storm events).
Some mesocosm and lab based studies use synthetic effluent with dissolved nutrients, which
may exaggerate treatment efficiencies, especially for P and metals which are usually bound
to particulates.
Thus, the main considerations when examining performance of a FTW are:
1. Dissolved oxygen: aerobic/anoxic/anaerobic. Natural aeration or artificial aeration
through bubblers. With aerobic basins tending to towards nitrification and anaerobic
basins tending towards denitrification.
2. Carbon sources: either naturally, through organic carbon, or added artificially, to
enhance denitrification rates. Decomposition of organic carbon also results in
increased ammonia production within the basin.
3. pH: with alkaline pH increasing nitrification and acidic pH increasing denitrification.
4. Root mass: aiding removal of particulates due to physical filtering and settling
processes
5. Mixing: circulation of water to aid the nutrient supply to microbiological processes.
6. Plug flow or continuous flow: affecting residence times and nutrient gradients.
7. Concentrations of inflow pollutants: with higher nutrient supply increasing rates of
decomposition/nitrification/denitrification unless limited by another factor.
8. Changes in the FTW chemistry with time. Often pH and redox potential drops due to
microbiological processes and restriction of oxygen diffusion from the surface.
Thus, direct comparison between different FTWs has little meaning, and the best
comparison is with a relevant control basin. This is often a basin without an island which is
receiving the same effluent, however sometimes it is before and after the retrofitting of an
island, which doesn’t guarantee exactly the same effluent inputs.
New treatment systems can take over a year to stabilise, and even then they can have high
variation in treatment efficiency, especially if environmental conditions vary or sinks (such as
sediment adsorption) become saturated. However, significantly higher performance of FTWs
can be noticed in as little as two days (Van de Moortel et al., 2010), particularly in relation to
Page 21 of 44You can also read
